curraiOpenTelemetry LLM tracing
Send OpenTelemetry traces to an LLM-aware dashboard
Currai ingests OTLP traces and turns model calls, tools, retrievers, tokens, cost, and latency into LLM observability views your team can use.
Keep your instrumentation, add LLM context
OpenTelemetry is the standard way to move traces across services. Currai accepts OTLP/HTTP so teams can send traces from existing OpenTelemetry exporters while getting views designed for LLM apps.
That means prompts, completions, model attributes, usage, latency, and nested spans remain connected in one trace.
- Export OTLP traces directly to Currai.
- Map GenAI semantic attributes into Currai generations and spans.
- Use OpenTelemetry from languages beyond Python and TypeScript.
For services that already use OTel
If your backend already has OpenTelemetry, Currai can receive LLM traces without forcing a separate tracing stack. Point the exporter at Currai and keep your service-level instrumentation strategy intact.
